Causal relationships of infection with Helicobacter pylori and herpesvirus on periodontitis: A Mendelian randomization study

BackgroundTo explore the causal association between Helicobacter pylori (H. pylori) infection, herpesvirus infection and periodontitis (PD) from a genetic perspective using Mendelian randomization (MR).MethodsThe PD data were derived from genome-wide association study (GWAS) from the Dental Endpoints (GLIDE) consortium, and the FinnGen Biobank provided data on H. pylori and herpesvirus infections. In addition, we examined GWAS data for subtypes of H. pylori and herpesvirus infection. Inverse variance weighting (IVW) was selected as a major analysis technique, and weighted median (WM), weighted model, simple model, and MR-Egger regression were added as supplementary methods. To verify the findings, the effects of pleiotropy and heterogeneity were assessed.ResultsGenetically predicted H. pylori infection (OR = 0.914, 95%CI = 0.693–1.205, P = 0.523), anti-H. pylori VacA (OR = 0.973, 95%CI = 0.895–1.057, P = 0.515), anti-H. pylori CagA (OR = 1.072, 95%CI = 0.986–1.164; P = 0.102), Epstein-Barr virus (EBV) infection (OR = 1.026, 95%CI = 0.940–1.120, P = 0.567), Herpes simplex virus (HSV) infection (OR = 0.962, 95%CI = 0.883–1.048, P = 0.372), cytomegalovirus (CMV) infection (OR = 1.025, 95%CI = 0.967–1.088, P = 0.415), EBV nuclear antigen-1 (EBNA1) (OR = 1.061, 95%CI = 0.930–1.209, P = 0.378), EBV virus capsid antigen (VCA) (OR = 1.043, 95CI% = 0.890–1.222, P = 0.603), HSV-1 (OR = 1.251, 95%CI = 0.782–2.001, P = 0.351), HSV-2 (OR = 1.020, 95%CI = 0.950–1.096, P = 0.585), CMV IgG (OR = 0.990, 95CI% = 0.882–1.111, P = 0.861) were not associated with PD, indicated that H. pylori and herpesvirus infection had no causal relationship to PD. Reverse studies also found no cause effect of PD on H. pylori or herpesvirus infection. The results of the sensitivity analysis suggested the robustness of the MR results.ConclusionThis study offered preliminary proof that H. pylori and herpesvirus infections were not causally linked to PD, and vice versa. However, more robust instrumental variables (IVs) and larger samples of GWAS data were necessary for further MR analysis.